Teacher Shocks and Student Learning: Evidence from Zambia.∗

We examine the effect of shocks to teacher inputs on child performance in school. We start with a household optimization framework where parents spend optimally in response to teacher and other school inputs. This helps to isolate the impact of shocks to teacher inputs. As a proxy measure for these shocks, we use teacher absenteeism during a 30 day period. Shocks to teacher inputs have a significant impact on learning gains. In a sample of students who remained with the same teacher over the two years for which we have test score data, shocks associated with a 5 % increase in the absence rate reduced learning by 4-8 percent of the average gains during the year. The size and precision of these estimates is identical for both Mathematics and English. We document that health problems account for over 60 percent of time spent in absence–this is not surprising in a country deeply affected by the HIV/AIDS epidemic. Tackling health problems of teachers and/or reducing the impact of absences by increasing the public provision of teachers (allowing for substitute teachers) is likely to have positive impacts on learning. ∗Corresponding Author: James Habyariman